Methods and devices for predicting dimerization in nucleic acid amplification reaction

Abstract

Proposed is a computer-implemented method for predicting a dimerization in a nucleic acid amplification reaction. The method may include accessing a dimer prediction model learned by a transfer learning method, and providing an input data to the dimer prediction model. The input data may include a sequence data of an oligonucleotide. The method may also include obtaining a prediction result for the dimerization of the oligonucleotide from the dimer prediction model.

What is claimed is: 
     
         1 . A computer-implemented method for predicting a dimerization in a nucleic acid amplification reaction, comprising:
 accessing a dimer prediction model learned by a transfer learning method;   providing an input data to the dimer prediction model, wherein the input data comprise a sequence data of an oligonucleotide; and   obtaining a prediction result for the dimerization of the oligonucleotide from the dimer prediction model.   
     
     
         2 . The computer-implemented method of  claim 1 , wherein the oligonucleotide comprises a primer. 
     
     
         3 . The computer-implemented method of  claim 2 , wherein the oligonucleotide comprises a forward primer and a reverse primer. 
     
     
         4 . The computer-implemented method of  claim 1 , wherein the dimerization comprises at least one selected from the group consisting of (i) a dimerization formed between two or more oligonucleotides and (ii) a dimerization formed in one oligonucleotide. 
     
     
         5 . The computer-implemented method of  claim 1 , wherein the dimer prediction model is a model obtained by fine-tuning a pre-trained model. 
     
     
         6 . The computer-implemented method of  claim 5 , wherein the pre-trained model uses a plurality of nucleic acid sequences as a training data. 
     
     
         7 . The computer-implemented method of  claim 6 , wherein the plurality of nucleic acid sequences are obtained from a specific group of an organism. 
     
     
         8 . The computer-implemented method of  claim 6 , wherein the pre-trained model is trained by a semi-supervised learning method performed in which a mask to some of bases in the nucleic acid sequences is applied and then an answer of the masked base is found. 
     
     
         9 . The computer-implemented method of  claim 8 , wherein the pre-trained model is trained by using nucleic acid sequences tokenized with tokens each having two or more bases. 
     
     
         10 . The computer-implemented method of  claim 9 , wherein the tokens comprise each bases tokenized by (i) dividing the nucleic acid sequences by k unit (wherein k is a natural number) or (ii) dividing the nucleic acid sequences by a function unit. 
     
     
         11 . The computer-implemented method of  claim 5 , wherein the fine-tuning is performed using a plurality of training data sets, each training data set comprises (i) a training input data comprising a sequence data of two or more oligonucleotides and (ii) a training answer data comprising a label data as to occurrence and/or non-occurrence of dimer of the two or more oligonucleotides. 
     
     
         12 . The computer-implemented method of  claim 11 , wherein the fine-tuning comprises (i) joining sequences of the two or more oligonucleotides by using a discrimination token and (ii) tokenizing the joined sequences to obtain a plurality of tokens. 
     
     
         13 . The computer-implemented method of  claim 12 , wherein the fine-tuning comprises (i) predicting a dimerization of the two or more oligonucleotides using a context vector generated from the plurality of tokens, and (ii) training the pre-trained model for reducing a difference between the predicted result and the training answer data. 
     
     
         14 . The computer-implemented method of  claim 5 , wherein the dimer prediction model comprises a plurality of models generated by fine-tuning the pre-trained model in accordance with each of reaction conditions used in the nucleic acid amplification reaction. 
     
     
         15 . The computer-implemented method of  claim 14 , wherein obtaining the prediction result comprises obtaining a plurality of prediction results for the dimerization from the plurality of models or obtaining a prediction result from a model corresponding to a reaction condition matched to the input data among the plurality of models. 
     
     
         16 . The computer-implemented method of  claim 11 , wherein the training input data further comprises a data of a reaction condition used in the nucleic acid amplification reaction, and the dimer prediction model comprises one model generated by fine-tuning the pre-trained model using the plurality of training data sets. 
     
     
         17 . The computer-implemented method of  claim 16 , wherein the input data further comprises a data of a reaction condition, whereby the prediction result for the dimerization is obtained based on the sequence data and the data of the reaction condition. 
     
     
         18 . The computer-implemented method of  claim 14 , wherein the reaction condition is a reaction medium, a reaction temperature and/or a reaction time used in the nucleic acid amplification reaction. 
     
     
         19 . A computer-readable recording medium storing a computer program including instructions that, when executed by one or more processors, enable the one or more processors to perform a method for predicting a dimerization in a nucleic acid amplification reaction by a computer device,
 the method comprising:
 accessing a dimer prediction model learned by a transfer learning method; 
 providing an input data to the dimer prediction model, wherein the input data comprise a sequence data of an oligonucleotide; and 
   obtaining a prediction result for the dimerization of the oligonucleotide from the dimer prediction model.   
     
     
         20 . A computer device for predicting a dimerization in a nucleic acid amplification reaction, the computer device comprising:
 a processor; and   a memory that stores one or more instructions that, when executed by the processor, cause the computer device to perform operations, the operations comprising:
 accessing a dimer prediction model learned by a transfer learning method; 
 providing an input data to the dimer prediction model, wherein the input data comprise a sequence data of an oligonucleotide; and 
   obtaining a prediction result for the dimerization of the oligonucleotide from the dimer prediction model.